Critical heat flux (CHF) tests were performed at low pressure in a close-packed rod bundle. The electrically heated test bundle had geometrical configurations the same as those of the Power Burst Facility nuclear core. Existing low-pressure CHF correlations, namely, those of Bernath and Lund, did not correlate well with the test data. The Bernath correlation overpredicts CHF in some cases by a factor of 5 when compared with measured values. Lund’s correlation overpredicts CHF at measured CHF values above 1.5 MW/m2, and underpredicts CHF at measured CHF values below 1.5 MW/m2. These CHF tests provided the first close-packed rod bundle data with a sufficient data base to develop a correlation. The study examined CHF with absolute coolant system pressures of 117 to 255 kPa, mass velocities of 1992 to 4830 kg/s· m2, and subcooling of up to 53°C, with a rod spacing of 1.02 mm. The effect of rod bowing was examined with the rod spacing reduced in varying degrees to a minimum of 0.0508 mm. Motion pictures of the rod bundle during CHF with nominal spacing and bowed rods show that CHF occurs in the rod gap and does not propagate azimuthally on the rod surface. A CHF correlation developed from the test data correlates the data with a standard deviation of 8.79%.